Connected Pathology in the UK: Part 2

Continuing our series of interviews with leaders of the UK’s National Pathology Imaging Co-operative (NPIC) team, we chatted to Basharat Hussain. Amongst the many challenges of such a massive digital pathology deployment, we addressed the two electronic elephants in the room: interoperability and data storage.

What’s your background and your current role on the NPIC team?
 

I have a degree in computing, and have worked in the public sector for 25 years, primarily in local government and the NHS. With 20-plus years of project and program management experience, I specialize in dealing with complex business change that requires innovative solutions.

At NPIC I’m director of digital transformation and AI deployment, and responsible for delivery of a national digital pathology solution, with artificial intelligence capabilities, to over thirty hospitals across the national health service (NHS). 

What does that role involve?
 

The diversity of the role means that there is always something going on. Typically my days are full of meetings with internal partners, finance procurement, and other key internal stakeholders like Leeds Teaching Hospitals executives. I also have strategic meetings with our partners and suppliers, such as NHS England, Office of Life Sciences, Innovate UK, and commercial partners. As well as that, I manage a team at NPIC and oversee their delivery objectives.

Often there are pressures of working towards tight deadlines and timescales, especially around funding applications. I juggle that with attending meetings and conferences around the country, not only to showcase our work and achievements, but also to gauge how the industry is evolving.

Across the NPIC program, what have been your highlights so far?
 

One of them was securing significant funding at the start of the program, as well as additional funding throughout the course of the program. Another was creating a highly motivated and capable team from scratch. Growing the team has been a significant highlight for me. 

On top of that, it’s been very satisfying to build the largest digital pathology PACS platform in the world for our NHS partners, which is capable of scaling to exabyte levels of storage. But it isn’t just about technology. The biggest achievement is that the platform is now being used in many hospitals today to improve patient care.

How are all those exabytes of data going to be stored?
 

This is one area that significantly impacts on the ability to deploy digital pathology at scale. The first question is how long do you keep the images? The current Royal College of Pathology guidelines state that you need to keep images for two pathology lab cycles, or about eight years. Those are the timescales we intend to follow on our digital pathology platform, which currently has over fifty petabytes of storage across two data centers, and is scalable to thousands of petabytes, or exabyte levels. We have a publicly funded private cloud solution to address the data storage challenge. 

For us, the technological challenge around storage isn’t the issue – it’s whether we have enough funding to keep growing it to the necessary levels. In total there are about 109 UK hospitals running histopathology services. We’re currently working with around thirty of them. In order to scale up the network for all of them, we’d be talking about thousands of petabytes of storage over 8 years. The plan is that as we grow the service, the income generated pays for the storage and the scaling up of the storage system. What we can’t control is the number of hospitals who join the program.

How are you addressing the interoperability challenge that comes up in digital pathology workflows?
 

Interoperability has been a challenge across the NHS. Our approach has been to use and evolve the DICOM standard. All our partners have agreed to sign up with our program and implement DICOM across the scanner platforms as well as the image management systems.

We also use HL7 messaging for a standardized approach when integrating systems. Fundamentally, I hope that taking a national approach to digital pathology with the platform that we’ve created means that there are less systems that need to be interoperable.

Of course, we’re not yet at the stage of addressing all interoperability issues, but we’re certainly making it easier to address by reducing the number of moving parts. 

What have the other main challenges been so far?
 

There has been significant investment in NHS pathology labs recently, which has led to broader laboratory transformation programs running in parallel with the NPIC rollout. One example of this is the laboratory reconfiguration project. Up and down the country, there has been a rationalization of laboratories in an attempt to achieve more standardization across regions, as well as the procurement and implementation of lab information management systems (LIMS). These dependencies have impacted our plans due to conflicting activity within the same space. Inevitably, any delays on those lab upgrades have directly impacted our timeframes.

Another challenge has been the variation in laboratory set ups. We often find that certain prerequisites, such as barcoding of slides, are lacking. And the final challenge has been to try to integrate legacy LIMS, which often have limited support and development capabilities. 

How far along are you with AI deployment?
 

The program is advancing its research into a number of AI tools, led by Daljeet Bansal. The research is quite advanced, but there is still quite a bit of work to do regarding AI deployment within a clinical setting at scale. We’re currently running some pilot projects to better understand how AI could be deployed to all of our NHS partners via our national platform.

Could you tell me about the multisite digital networks that are now operating, thanks to the NPIC program?
 

There are a number of NHS sites live and using the platform on a daily basis for primary diagnosis as well as second opinions and multidisciplinary meetings. The unique selling point of the solution is its ability to work across geographical boundaries and regions without the need for additional systems sitting alongside to enable integration and interoperability. This has the dual advantages of simplifying our IT and reducing the cost of deployment and management.

What lies ahead for the team?
 

For the next twelve to eighteen months, we’ll continue to deploy across the forty or so partner NHS trusts who have signed up for the program. We’ve also been inundated with new NHS hospitals interested in joining the program, so that will be an area of continuing focus for us.

The deployment of AI tools across the clinical system will be a priority. We hope to make significant progress with this in the coming year.

Finally, having a large national platform will also provide new opportunities and capabilities and I hope to leverage this to support other areas.